License Plate Recognition System Based on Artificial Intelligence with Different Approach

Abstract

An license plate recognition system (LPRS) generally provides control and security. These systems are created using methods such as artificial intelligence, machine learning, artificial neural networks (ANN), deep learning, fuzzy logic, expert systems, and image processing. This study aims to create an LPRS using artificial intelligence and image processing techniques. The prepared system is for rectangular-sized plates. An LPRS consists of 3 main stages. The first stage is to detect the plate region. At this stage, converting to grayscale, bilateral filtering, canny filtering, and contour were applied to vehicle images. The second stage is to crop the plate region. In the second stage, the masking method was employed. The pytesseract algorithm was used to recognize license plate characters in the last stage. To create the system, Raspberry Pi 4 Single-Board Computer (SBC) was used for hardware; python programming language was utilized for software. The results showed that the system worked successfully at the rate of 100% in the first two stages and at the rate of 91.82% in the last stage. The results suggest that the system works successfully.